NASA Technical Memorandum 83331

The Thermal Fatigue Resistance of H-13Die Steel for Aluminum Die Casting Dies

< N A S A - T . v - 8 3 3 3 I ) IH-£

R E S I S T A N C E O F H - 1 3 D I E S T E I L F C f A L U M I N U MC I I C A S I N G D I E S ( N A S A ) 2 2 p E C A 0 2 / M F A J 1

C S C L NF Dnc la sG3/26 ^ 2 C 9 7

Betty J. Barrow£ew/s Research CenterCleveland, Ohio

Prepared for theLet's Do It Ourselves: In Science and Technologysponsored by the National Technical AssociationBaltimore, Maryland, August 2-7, 1982

NASA

https://ntrs.nasa.gov/search.jsp?R=19830026832 2018-05-17T03:06:08+00:00Z

U.S. Department of CommerceNational Technical Information Service

N83-35103

THE THERMAL FATIGUE RESISTANCE OF H-13DIE STEEL FOR ALUMINUM DIE CASTING DIES

LEWIS RESEARCH CENTERCLEVELAND, OH

AUGUST 82

THE THERMAL FATIGUE RESISTANCE OF H-13 DIE STEEL

FOR ALUMINUM DIE CASTING DIES

Betty J. Barrow

National Aeronautics and Space AdministrationLewis Research CenterCleveland, Ohio 44135

ABSTRACT

This investigation was conducted to determine the effects of welding,selected surface coatings and stress relieving on the thermal fatigue resist-ance of H-13 Die Steel for aluminum die casting dies.

Eleven thermal fatigue specimens were utilized in this investigation andare categorized as follows: control, stress relieved, welded, and surfacetreated (five different surface treatments were evaluated). Stress relievingwas conducted after each 5000 cycle interval at 1050° F for 3 hours. Fourthermal fatigue specimens were welded with H-13 or maraging steel welding rodsat ambient and elevated temperatures and subsequently, subjected to differentpost-weld heat treatments. The surface treatments used are proprietary butare described in generic terms. Crack patterns were examined at 5000, 10 000,and 15 000 cycles. The thermal fatigue resistance is expressed by two crackparameters which are the average maximum crack and the average cracked area.

The results indicated that a significant improvement in thermal fatigueresistance over the control was obtained from the stress-relieving treatment.Small improvements in thermal fatigue resistance were obtained from the H-13welded specimens and from a salt bath nitrogen and carbon-surface treatmentwhich is attributed to the alleviation of the residual stresses and the forma-tion of a diffusion layer at the surface.. The other surface treatments andwelded specimens either did not affect or had a detrimental influence on thethermal fatigue properties of the H-13 die steel.

INTRODUCTION

Stress, Erosion and Corrosion Effects

In die casting, the molten metal is forced into the mold or die under highvelocity and pressures and held against the die wall until solidificationoccurs.1 After the casting has cooled sufficiently, the die is opened andthe part is ejected by pins which are mechanically activated. To increase therate of cooling of the die and casting, the dies are usually water cooled.

Dies for aluminum alloys are significantly higher in cost than those forlower melting temperature alloys. Depending on its intricacy, the cost of thedie may be more than the machinery required to operate it. Efforts to holdcost within reasonable limits must include increasing the production life ofthe die. Therefore, prevention of die failure can make a major contributiontoward the profitability of the die casting operation.

Aluminum die casting dies are subjected to high thermal-mechanical loads.As a result, the service life of these dies can be limited. The major causesof die failure are heat checking, cleavage or gross cracking, decarburization,erosion damage and indentations. Of these, the most frequent cause of die

failure is due to heat checking. Heat checking is caused by thermal fatiguewhich results from the rapidly alternating temperatures and thermal gradientsnormal to the die surface. These cyclical temperature fluctuations inducestress and strain combinations that can produce cracking of the materialresulting in premature failure.^ The magnitude of the thermal stresses andstrains depend upon the mechanical and thermophysical properties of the mater-ial, cooling rates, and the maximum and minimum temperatures of the thermalcycle.^

Thermal fatigue cracks or heat checks initiate as pits at points of maximumstress and at inclusions or imperfections present on or near the specimen sur-face. In ductile materials such as H-13 die steel, cracking is preceded bysurface markings that appear as small grooves. Heat checking occurs as a finenetwork of cracks located on the die surface. Propagation of these cracks isperpendicular to the die face. Heat checks are transgranular and appearstraight and sharp with a greater thickness near the surface. Figure 1 is aphotomicrograph of the control specimen at SOX (unetched) to show the appear-ance of a heat check on the die surface.

When molten liquid metal is injected into the cold die, the surface isimmediately subjected to a severe temperature increase. This temperature israpidly decreased by conduction of heat back into the underlying lower temper-ature die steel. The temperature variations at the die surface during a cast-ing cycle are illustrated in figure 2.4 The magnitude of this temperaturefluctuation is determined by the heat transfer characteristics of the die sur-face due to the presence of coatings, oxide layers and dimensional configura-tions. The die surface in immediate contact with the molten metal expands andcontracts continuously as the casting cycle is repeated, resulting in thermalstresses at the surface due to self-constraint. This continued sequence ofalternating stresses can result in plastic deformation of the material. Underthese conditions, the induced stresses are capable of exceeding the fatiguestrength of the material leading to heat checks.

Both the initiation and propagation of thermal fatigue cracks are furtherinfluenced by corrosion. Throughout the operation of the die, it is exposedto an oxidizing atmosphere. Progressive oxidation occurs with additionalcycling. As a result, corrosion pits which form are filled with oxides. Theoxides present occupy a larger volume than the metal which they replaced,resulting in a wedging action and a tri-axial tensile stress state at the tipof the crack. These tri-axial tensile stresses are a major contributor tocrack propagation.

The purpose of this investigation is to determine what effect selected sur-face treatments, various welding procedures, and stress-relieving has on thethermal fatigue resistance of H-13 die steel. A comparative study will deter-mine which surface treatment and welding procedure has the most beneficial andmost detrimental effect on the thermal fatigue properties of the H-13 diesteel. Finally, a metallographic examination of the transverse cross-sectionof each corner will show the interaction of the surface treatments and weldswith the base metal. This analysis is used to determine the relationshipbetween the thermal fatigue properties and microstructures.

MATERIALS AND PROCEDURES

The selection of the appropriate base metal to use for die casting diesrequires a consideration of its thermal fatigue resistance, toughness, hothardness, machinability and cost. H-13 die steel is most widely used in thedie casting industry because it offers the best compromise of this range of

material characteristics. Also, H-13 die steel possesses a low coefficient ofthermal expansion and a high thermal conductivity which are essential for goodthermal fatigue resistance. The optimum hardness level f<jr the maximum ther-mal fatigue resistance in H-13 die steel is Rockwell C 46^ (Rc). A hard-ness level over RC46 lowers the toughness and can lead to gross cracking.It the hardness level is lower than RC46, the H-13 die steel will becomesusceptible to heat checking and /or erosion.

The eleven thermal fatigue specimens utilized in this investigation weremachined from a 3 1/2 inch diameter bar of H-13 die steel to meet the dimen-sions shown in figure 3 plus a 0.050 inch allowance for grinding after heattreatment to reduce surface decarburization. The composition of this steelcan be seen in table 1. After machining, all specimens were austenitized at1850° F for 2 hours in a salt bath and then oil quenched. ^Following thequench, the specimens were double tempered in salt at 1100° F for 1 hour andthen air-cooled, the hardness at this time was 48-49 Rc, so the thermalfatigue test specimens were tempered at 1100° F for 7 1/2 hours in an air fur-nace. This lowered the hardness to the required value of R .46.

In all specimens where a surface coating was to be applied, the two cornersto be coated (alternate corners) were finished to final dimensions. The 0.050inch stock was not removed from the other alternate corners which were to beused as the control. This 0.050 inch of stock which remained was removed as alast operation before testing to eliminate by-products of the coating treat-ment from the control corner surfaces.

The coatings tested were proprietary. The commercial designation of thesecoatings is not employed but each are described in generic terms. Coating Ais a salt bath nitrogen and carbon diffused surface treatment. Coating A wasapplied at 1060° F for 2 hours. Coating B is a low temperature salt bathtreatment. Coating C is a Boron diffusion surface treatment. This processtempered the steel; therefore, it was necessary to re-heat treat the specimenby austenitizing at 1950° F for 2 hours. The process was completed by adouble-temper at 1100° F for 1 hour in order to maintain the hardness value ofRC46. Coatings D and E are nitrogen diffused surface treatments. Coating Dwas applied with an ionized gas at 986° F for 24 hours and Coating E was gasnitrided at 1000° F for 24 hours.

The specimen to be stress-relieved was heated to 1050° F for 3 hours andallowed to furnace cool to 800° F after every 5000 cycles. The specimen wasthen air-cooled to room temperature. The corners of the fatigue specimens tobe welded were undercut along the entire length of the specimen a distance of3/8 inch from the edge to a depth of 0.094 inch. All welds were applied withthe gas shield metal ARC welding process (GMAW). The compositions of the mar-aging and H-13 welding rods are listed in table I. The weld temperature andpost-weld heat treatments are listed in table II.

Metallographic examination of these weld areas indicated that the H-13welds were made in six passes and the maraging steel in ten passes. The welddeposits were made sequentially with no waiting period to reach the specifiedwelding temperature before depositing the next weld bead. This procedureundoubtedly resulted in higher interpass temperatures than the stated weldingtemperatures. The .010" radius was placed on all corners before the post-weldheat treatment.

A schematic diagram of the thermal fatigue test equipment is shown in fig-ure 4. The function of this equipment was to simulate the environmental con-ditions indicative of die casting. The specimen was immersed into a molten380 aluminum alloy bath at an average temperature of 1300* F for 13 seconds,followed by a 23 second cooling period in air at ambient temperatures.

The specimen was internally cooled by a continuous flow of water at aerateof 1 gallon per minute. To maintain the molten bath temperature of 1300° F,two chromel-alumel thermocouples were employed to monitor and record this tem-perature throughout cycling. Immediately prior to the immersion of the speci-men into the bath, its surface was sprayed with a water-base die lubricant.The lubricant mixture was a diluted solution consisting of die lubricant,^ChemTrend F-3-L and water at a ratio of 1 to 50. The specimen was rotated 90*every 3750 cycles to equalize any variation in the density of the lubricantspray on each corner. Rapid cooling from an over abundance of lubricant canhave a detrimental effect on thermal fatigue resistance because it increasesthe temperature differential.

The surface of each specimen was examined every 5000 cycles until the testwas complete at 15 000 cycles. The oxide layer which accumulated over thecorners to be investigated was removed by manually polishing the specimen witha V-shaped die block utilizing 240 grit and 400 grit grinding papers. Onlythose cracks, which initiated along the edge of the corner and located withinthe central 3 inches of the specimen, were counted. The measurements weretaken at a magnification of 100X with a Leitz microhardness testing unit.

The length of each crack was used to describe the thermal fatigue resist-ance in terms of two crack parameters. They are the average maximum cracklength and the average cracked area. The average maximum crack length is themeasurement of the longest crack length which initiated from the edge of thecorner on each corner divided by the number of corners. This parameter isreferred to by the symbol d. The average cracked area is the sum of the num-ber of cracks on each corner multiplied by the square of the crack length anddivided by the number of corners. This parameter is referred to by the symbolind^. These parameters are calculated for each specimen. Both parametersare significant in determining the thermal fatigue resistance of the diebecause the longest crack within the die surface and the extent of crackingcan be determined by the average maximum crack length and the cracked arearespectively. The values of these crack parameters are indicative of theservice life of the die. Figure 5 is a photograph of the area to be investi-gated. The dark surface is the thermal fatigue crack. The light region isthe surface area of the specimen which was cracked open to expose the thermalfatigue crack. Figure 6 is a scanning electron microscope photomicrograph at20X which compares the two exposed surfaces of the heat check corner. Theright side of the photomicrograph is the surface of the thermal fatiguecrack. This smooth, cleavage surface is indicative of a brittle transgranularfailure. In comparison, the rougher surface on the left side of the photo-micrograph contains more dimpled regions which is indicative of a ductilefracture.

Microhardness readings were taken along the diagonal of the transversecross-section every 50 microns beginning at the edge of the corner, utilizing500 gram load in a Leitz microhardness testing unit. The hardness measure-ments were taken to determine any hardening or tempering of the base materialfrom the special surface treatments. These measurements were used to deter-mine the relationship between the tabulated crack parameters and the micro-structural changes.

DISCUSSION OF RESULTS

Thermal Fatigue Properties

The average cracked area (znd^) and average maximum crack length (d) pa-rameters used to determine the heat checking resistance throughout the three

5000 cycle test intervals have been tabulated in tables III and IV for thecoated and welded specimens respectively. By using these quantitative meas-urements, table V (a and b) was developed to arrange the specimens in order ofdecreasing thermal fatigue resistance. Figures 7 and 8 are plots of the aver-age cracked area and maximum crack length, respectively. A steeper slope ofthe lines in figures 7 and 8 for each specimen indicates greater thermalfatigue cracking and can be used to predict initiation and propagation ratesqualitatively. From the ranking of the specimens in table V and the intersec-tion of lines in figures 7 and 8 it can be seen that the two crack parametersare independent of one another by the varying positions of the specimensthroughout testing.

Table V and figures 7 and 8 at 15 000 cycles show that the stress relievedspecimen demonstrated the best overall thermal fatigue resistance. Theimproved thermal fatigue resistance of the stress relieved specimen over thecontrol specimen is attributed to the removal of residual stresses which weregenerated during the thermal cycle. Another specimen which exhibited lesscracked area compared to the control and possessed superior thermal fatigueresistance with respect to all the other coatings was the Specimen withCoating A.

The bath for Coating A was composed primarily of cyanide and cyanate com-pounds. The treatment in the salt bath caused the diffusion of carbon andnitrogen into the H-13 die steel. Since nitrogen is more soluble than carbonin ferrous metals, it diffuses more readily into the metal. However, the car-bon formed iron-carbide particles at or near the surface of the corner. Theseparticles act as nuclei, precipitating some of the diffused nitrogen to form atough compound zone of carbon-bearing, epsilon nitride. This compound isrelatively thin (= .005 inch) as formed in a normal 90 minute treatment. Noneof the brittle Fe£N phase was formed.6 With more nitrogen available inthe bath than can be absorbed and held at the surface in the compound zone,the nitrogen begins to diffuse further into the matrix of the H-13 die steel.It should be noted that the diffusion zone is relatively limited. Theincrease in thermal fatigue strength is attributed to this diffusion zone, notthe compressive forces created at the surface.^

The coating which demonstrated the worst thermal fatigue resistance wasCoating E. The microstructure of the transverse cross-section of this speci-men magnified at 150X can be seen in figure 9. It consists of various diffu-sion layers which result from the nitriding process. The nitriding processcan produce a considerable amount of iron nitrides in the surface layers whichresults in increased hardness and can lead to surface brittleness. Apparent-ly, the surface brittleness contributed to inferior thermal fatigue propertiesof this specimen. From figures 7 and 8 it can be seen that Coating E demon-strated a large amount of cracking up to 5000 cycles. However, the crackpropagation rate of this specimen decreased considerably after 5000 cycles.This is probably due to the fact that the crack depths exceeded the hardenedcoating thickness on the surface of the specimen.

Two specimens identified as maraging steel weld I and II were tested inthis investigation. The Marage I specimen contained a considerable amount ofporosity, whereas the Marage II specimen was free from these defects. It isapparent from the location of the heat checks through this porosity in theMarage I welded specimen that these defects made a major contribution to thepoor thermal fatigue resistance of the weld. Figure 10 is a photomicrographtaken at 45X of a gas pore located in the Marage I welded corner. Gas poresact as stress risers and the longest heat checks which occurred in this cornerinitiated at these defects. This behavior of the Marage I Weld is demon-strated in figures 7 and 8. The Marage II specimen was produced with a

porosity-free weld. However, the Marage II welded corner exhibited poorerthermal fatigue resistance than the Marage I welded corner. The poor thermalfatigue behavior of the Marage II Weld is attributed to the presence of con-siderable amounts of reverted austenite which formed in the segregated por-tions of the weld that have higher nickel, carbon,^nitrogen, and manganesecontents. This austenite formed in the 1075e-1100° F temperature range towhich the specimen corners were heated during immersion. Austenite has poorthermal fatigue resistance because of its low strength, high thermal expan-sion, and poor thermal conductivity. The microstructure of the Marage II weldis illustrated in figure 11. This microstructure shows a background of graycolored martensite and the light regions are indicative of the segregatedareas that have reverted to austenite.

The H-13 welds were all free of porosity and when tested, possessed goodthermal fatigue properties. The corners welded with H-13 welding rods hadthermal fatigue resistance similar to one another irrespective of its weldingprocedure (see table II). Because these results were so similar, the averageof crack parameters were calculated after 5000, 10 000 and 15 000 cylces andplotted in figures 7 and 8. It is apparent from these figures that the aver-age maximum crack length and crack area in these welded corners are similar toand perhaps slightly better than the control after 15 000 cycles. The goodperformance of these welds may be attributed to their somewhat greater hard-ness values over the heat treated control corners, as will be demonstratedlater in this paper.

In comparing the results of the various welded H-13 corners, it appearsthat the post-weld heat treatment at 900° F for^S hours performed slightlybetter than the post-weld heat treatment of 400° F for 3 hours. The highertemperature may have relieved more of the residual stresses in the weld beforetesting. Since the corners are heated to 1075e-1100e F at their highest tem-perature point during the thermal cycle, the effect of this post-weld treat-ment would be eliminated very quickly during thermal fatigue testing.

Microhardness Measurements

The microhardness readings were taken along the diagonal of the corner ofthe transverse metallographic specimen from the corner inward toward the basemetal. The first indentation was made as close as possible to the corneredge, with the remaining readings taken at 50 micron increments. The resultsof the microhardness measurements on each thermal fatigue specimens are plot-ted in figure 12.

Significant softening of the corner with respect to the base metal isapparent to a depth of about 1200 micron along the diagonal of every specimenexcept the specimens with coatings D and E. The hardness of the heat treatedspecimen was OPH 460 or Re 46.5 before cycling and prior to the application ofthe coatings. The various coating treatments significantly increased thishardness into the 60 Rockwell C range but all except the specimens with coat-ings D and E were softened by thermal cycling. The poorest thermal fatigueresistance (figures 7 and 8) was exhibited by specimens 0 and E. The poorthermal fatigue resistance of these specimens is attributed to the presence ofthe brittle surface layer at the corner.

Attention is also called to the generally lower hardness of the specimenthat was welded with maraging steel and higher hardness of the corner of thespecimen welded with the H-13 welding rod. The microhardness values of theweld and the corresponding heat-affected zones of the Marage II and H-13welded specimens are plotted in more detail in figure 13. The welds were made

in six and ten welding passes, for the H-13 and Marage II weld, respectively.Therefore, the hardness varies for each weld depending on the thermal historyof the metal. Both welds eventually reach the base H-13 steel and have ameasurable, similar hardness in the 460 DPH range. However, the hardness ofthe maraging steel is significantly lower across nearly the entire weld areathan the H-13 steel. This lower hardness results from the presence of thereverted austenite in this steel. This austenite can be formed either fromthe high temperature portion of the thermal fatigue cycling or from the heataffect of subsequent welding passes. The hardness of the H-13 steel weldshows considerable variation depending on location across the weld deposit.The weld metal near the corner of the thermal fatigue specimen has been soft-ened by the thermal cycling of the test. The portion of the weld toward thebase metal has been hardened by subsequent weld deposits in this multipassweld. These hardened areas have not been softened by thermal cycling becausethe temperatures away from the corner do not reach levels as high as that ofthe corners. In some cases the H-13 weld material from earlier deposits hasbeen retempered by subsequent deposits accounting for the lower hardness.This different thermal history has produced a wide variation in hardness read-ings over the H-13 weld. The general hardness level is significantly higher,however, than for the Maraging II Weld. At some locations this microhardnessattains high levels and the weld is undoubtedly brittle at these locations.However, no cracks were noted and this high hardness apparently did not reducethe thermal fatigue resistance. A total of six weld beads were counted in theportion of the weld which still remained; this number of weld depositsaccounts for the variation in hardness values.

The stress-relieved specimen also had a somewhat higher microhardness thanthe control or corners with coatings A, B and C. The increased hardness ofthe stressed relieved specimen is caused by secondary hardening which is theresult of heat treating at 1050° F for 3 hours. This treatment may also havereduced the softening in this specimen by relieving the residual stressespresent. The increase in hardness may have contributed to the improved ther-mal fatigue resistance of the stress-relieved specimen over the coated speci-mens. However, the hardness value of Re 36 is still well below the originalhardness of the H-13 die steel before stress relieving.

CONCLUSIONS

The results were based on the two crack parameters called the average maxi-mum crack length and the average cracked area. Both parameters were calcu-lated in terms of the length of the heat checks present on the specimensurface. Each parameter is evaluated independently and is significant indetermining the service life of the die. The average maximum crack lengthindicates the length to which a heat check has propagated within the surface.The average crack area is indicative of magnitude of cracking throughout thecrack surface.

The following conclusions were reached in this investigation to determinethe influence of selected surface coatings, stress-relieving and various typesof welds on the thermal fatigue resistance of H-13 die steel.

1. The best thermal fatigue resistance was obtained with the specimen thatwas stress relieved at 1050° F after 5000 and 10 000 thermal fatiguecycles.

2. The thermal fatigue resistance of the die steel was slightly improved bycoating A, which is a proprietary liquid salt cyanide treatment.

3. The ion and gas nitriding coatings D and E reduced the thermal fatigueresistance sharply while the cracking occurred within the hardened surfacelayer.

4. The thermal fatigue resistance of the corners produced with H-13 elec-trodes was generally excellent. On the average, the thermal fatigue prop-erties of the H-13 welded specimens were slightly better than the controlspecimens (see figures 7 and 8). The best thermal fatigue resistance ofthe H-13 welded specimens was obtained when the 900° F post-weld heattreatment was employed. The higher hardness and finer structure of H-13welds as compared to the H-13 die steel may have produced the improvedthermal fatigue resistance.

5. The maraging welds were very poor in thermal fatigue resistance. In thecase of the Marage I weld, considerable porosity lowered the thermalfatigue resistance. Even though the second marage weld was sound, it alsopossessed poorer thermal fatigue properties than the Marage I weld. Theinferior thermal fatigue properties of this weld can be attributed to thepartial reversion to austentite in the segregated areas. The reversion toaustenite may have occurred because of the heat effect of subsequent weld-ing passes and the high temperature portion of the thermal fatigue cycle.The presence of austenite is detrimental to the thermal fatigue resistanceof H-13 die steel because of its reduced hardness, poor thermal conductiv-ity and high thermal expansion coefficient.

REFERENCES

1. Lindberg, R. A.: Processes and Materials of Manufacture. Second ed.,Allyn and Bacon, Inc., 1977, p. 294.

2. Gibbons, C. L.; Moore, A. A.; Young, R. M.: Thermal Fatigue of DieSteels: A Brief Summary and Perspective. Ind. Heat., vol. 48, no. 7,July 1981, pp. 24-29.

3. Gorbach, V. G.; Alekhin, V. G.; Kurgannova, G. L.: Determining ThermalFatigue of Steels for Die Casting of Aluminum Alloys. Met. Sci. HeatTreat. Met., vol. 19, no. 11, Nov. 1977, pp. 982-985.

4. Wallace, J. F.: Thermal Conditions in the Die. Foundary, vol. 96, no. 10,Oct. 1968, pp. 176-179.

5. Nagy, A.: Maraging Steels for Die Casting. Met. Prog., vol. 97, no. 4,Apr. 1970, pp. 70-71.

6. Taylor, E.: Tufftride: Only Skin Deep. Met. Eng. Q., vol. 11, no. 2,May. 1971, pp. 12-16.

TABLE I. - CHEMICAL COMPOSITIONS OF H-13 DIE STEEL,

H-13 WELDING ROD AND MARAGE WELDING ROD

[Compositions in weight percent.]

Element

C

Mn

Si

Ni

Cr

Mo

V

Co

Ti

Al

S

P

H-13 Die steel

0.38

.38

1.12

—

5.06

1.26

.94

—

—

—

.008

.014

H-13 Welding rod

0.39

.31

1.16

—

5.11

1.50

1.02

—

—

—.003

.014

Marage rod

0.007

.04

.04

18.53

—4.84

—7.22

.41

.10

.002

.002

TABLE II. - WELD PROCEDURES AND POST-WELD HEAT TREATMENTS

Specimen

BB

BC

BD

BE

Corner

A

C

B

D

B

D

B

D

Weldmaterial

H-13

H-13

Marage

H-13

H-13

Marage

H-13

H-13

Weldingtemgerature,

600

RT~80

Ambient

Ambient

Ambient

Ambient

600

RT-80

Post weldtreatment,

°F

900

900

900

900

900

900

400

400

Time,hrs

3

3

3

3

3

3

1

1

ORIGINAL PAGt ?§OF POOR QUALITY

10

TABLE III. - THERMAL FATIGUE TEST RESULTS OF H-13 DIE STEEL

WITH VARIOUS DIFFUSION COATINGS

Description

Control

Coating AControl

Coating BControl

Coating CControl

Coating DControl

Coating EControl

Corners

A,B,C,D

A,CB,D

A,CB,D

A,CB,D

A,CB,D

A,CB,D

Average maximum crack length:d(v X 102)

Number of cycles

5000

1.5

5.58.0

4.03.5

2.07.0

42.53.5

48.52.0

10 000

7.0

8.012.0

6.010.5

12.08.0

43.510.5

51.06.0

15 000

19.2

21.020.0

19.022.0

32.022.0

44.522.5

57.015.0

Cracked area:znd2(y2 x 106)

Number of cycles

5000

0.30

2.0614.97

0.201.23

0.1116.97

163.320.58

286.460.09

10 000

6.11

3.0826.83

5.5524.48

40.028.0

239.2510.30

378.4018.60

15 000

42.23

13.7353.71

57.7573.03

106.0554.51

276.5037.76

514.5044.48

.'J L f-;:r.T ^POOR QUALITY

11

TABLE IV. - THERMAL FATIGUE TEST RESULTS OF H-13 DIE STEEL WITH MARAGE WELD,

H-13 WELD, AND STRESS RELIEVED CORNERS

Specimen

BBa

BCa

BD3

BEb

BI

Corners

ACB,D

BDA,C

BDA,C

BDA,C

A,B,C,D

Description

H-13 Weld at 600°FH-13 Weld at 80°FControl

H-13 Weld at RTMarage -1 at RT (avg)Control

H-13 Weld at RTMarage -2 at RTControl

H-13 Weld at 600°FH-13 Weld at RTControl

Stress relievedH-13 Weld average

Average maximum crack length:d(P x 10

2)

Number of cycles

5000

5.04.03.5

3.020.03.0

6.012.08.0

5.06.05.0

3.04.8

10 000

8.05.04.5

6.533.07.0

16.047.011.5

6.08.06.5

3.58.25

15 000

11.09.07.5

14.564.015.5

23.079.015.5

16.013.010.5

6.214.17

Cracked area:£nd2(y2 x 106)

Number of cycles

5000

8.083.222.92

2.1115.332.95

4.6320.4711.03

4.884.735.21

2.134.60

10 000

16.5611.325.46

7.3147.586.44

28.30120.3233.30

6.449.569.02

3.3613.24

15 000

24.3215.1724.45

37.59135.5140.65

54.56258.3646.55

16.1624.4325.26

6.3428.70

aPost Weld Treatment:bPost Weld Treatment:

900°F for 3 hours.400°F for 1 hour.

ORIGINAL PAGE !SOF POOR QUALITY

12

TABLE V. - THE ARRANGEMENT OF FATIGUE SPECIMENS IN ORDER OF

DECREASING THERMAL FATIGUE RESISTANCE

(a) Average cracked area: znd2 106)

5000 Cycles 10 000 Cycles 15 000 Cycles

Coating CCoating BControlCoating AStress relievedH-13 Weld averageMarage-1Marage-2Coating DCoating E

Coating AStress relievedCoating BControlH-13 Weld averageCoating CMarage-1Marage-2Coating DCoating E

Stress relievedCoating AH-13 Weld averageControlCoating BCoating CMarage-1Marage-2Coating DCoating E

(b) Average maximum crack length: d (y x 102)

5000 Cycles

ControlCoating CStress relievedCoating BH-13 Weld averageCoating AMarage-2Marage-1Coating DCoating E

10 000 Cycles

Stress relievedCoating BControlCoating AH-13 Weld averageCoating CMarage-1Coating DMarage-2Coating E

15 000 Cycles

Stress relievedH-13 Weld averageCoating BControlCoating ACoating CCoating DCoating EMarage-1Marage-2

QR'GSkAL ^AS :OF POOR QUALITY

13

<* POOR

-

-

Figure 1. • The longitudinal cross-section of the control corner. Cracks initiatefrom the edge and propagate perpendicular to the edge of corner (Mag. 50X).No etch.

'max

E

'min

START OFMETAL FLOW

-CASTING SOLIDIFICATION PERIODl'MAXIMUM SURFACE TEMPERATURE

CASTING COOLINGPERIOD IN DIE

^LIQUIDLUBRICATION

rBASE DIETEMPERATURE

,'"DIE SUBSURFACETEMPERATURE

PERIOD OF ONE CASTING CYCLE

TIME

Figure 2. - An idealized temperature cycle during the die casting process.

ORIGINAL PAGE 18OF POOR QUALITY

HOLE MUST BECENTER + 0.005 in. ^

— \

~T2 + 0.003 in.

0.001 in.1

2 in.

' 1

\~HF

7 + 0.

1.5 + 0.010 in.

010 in.

I

<

1~ t

\

'/

X-ARE' FAT

,/^lw

r°

r ALL CORNERS MUSTBE SQUARE+ 0.003 in.

1-1/2 DRILL TODEPTH SHOWN1-1/4—11-1/2 NPT TAP

1-1/8 DEEP

AREA OF THERMALFATIGURE CRACKING

0.5+ 0.015 in.

t

Figure 3. - Thermal fatigue specimen.

<sAIRJ

CHROMEL-ALUMELCONTROL vTHERMO- \COUPLE —

n =

OUT-J

SOLENOIDVALVE

~"

LUBRICANT c

• — — • L-

"^PULLEYS—"

rWATER

^SPECIMEN

' r LUBRICANT; / SPRAY

j / -.T^A05

* i \' V\/ v\.-^ _X^

tlJ

-AIRCYLINDER

1"""^ i 1 _. T=l|', MOLTEN /

380 / /» ALUMINUM /

\ /A_ — J

yL 1 _

GASFURNACE

^-SiC Ci\uv«IDLC ur ic? rvuuv unrnun I

Figure 4. - Thermal fatigue test equipment.

ORIGINAL PAGE ISOF POOR QUALITY

Figure 5. - The exposed surface of a heat check found in the marage II weldedcorner. The darkened areas is the crack and the light area is the H-13 diesteel (Mag. 8X).

Figure 6. - The scanning electron photomicrograph of the surface of a heatcheck. The dark area on the right is a brittle failure associated with thethermal fatigue crack. The dimpled area on the left is the surface of theH-13 die steel exposed by mechanical failure (Mag. 20X).

500

400

300

200

100

ORIGINAL P,*.Q£ S3OF POOR QUALITY

COATING D

'MARAGE 2

MARAGE 1COATING C

1 ^

60 -

COATING E 40 -

ENLARGEMENT

10 15x10^NUMBER OF CYCLES

COATNG B

CONTROL

H-13WELDAVERAGE

COAT ING A

STRESSRaiEVED

ISxlO3

Figure 7. - Thermal fatigue test results - sums of squared cracked

lengths.

SOxlO2

g

MARAGE 2

ENLARGEMENT

COATING C

COATING A

CONTROLCOATING B

H-13WaOAVERAGE

STRESSRELIEVED

NUMBER OF CYCLES

Figure 8. - Thermal fatigue test results - average maximum cracklength.

ORIGINAL PAGE 19OF POOR QUALITY

Figure 9. • The transverse cross-section of the specimen with applied coating E.It consists of various diffusion layers which resulted from the nitriding processthe structure was etched by 2% nital solution (Mag. 150X).

Figure 10. • The photomicrograph of a gas pore found in the marage I weld. Thecavity acted as a stress riser which aided in the initiation and propagation ofheat checks. No etching solution was used (Mag. 45X).

OF POOR QUALITY

Figure 11. - The transverse cross-section of the marage II weld.Note the visible banding and segregated areas. The structurewas etched with 2% nital solution (Mag. 500X).

800

600

400

200

D COATING AO COATING B£ COATING CO COATING DA COATING E& CONTROLD H-13(900°F)0 STRESS RELIEVEDfc, MARAGE (900° F)

J I I200 400 600 800 1000 1200 1400

DISTANCE ALONG DIAGONAL, MICRONS2000 4000

Figure 12. - DPH hardness readings along diagonal of transverse metal-lography specimens.

ORIGINAL PAGE! rgOF POOR QUALITY

O H-13WELD700,— D MARAGE

400

300

200 I I12 16 20 24 28 32 36 40 44X102

DISTANCE ALONG DIAGONAL, p

Figure 13. - DPH hardness readings along diagonal of transverse metal-lographic specimens to illustrate varying hardnesses throughout theheat affected zones of welded specimens.

\

1. Report No. 2. Government Accenion No.

NASA TM-833314. Titlt and Subtitle

THE THERMAL FATIGUE RESISTANCE OF H-13 DIE STEELFOR ALUMINUM DIE CASTING DIES

7. Authorl i l

Betty J. Barrow

9. Performing Organization Name and Address

National Aeronautics and Space AdministrationLewis Research CenterCleveland, Ohio 44135

12. Sponsor mg Agency Name «nd Address

National Aeronautics and Space AdministrationWashington, D. C. 20546

3. Recipient's Catalog No.

5. Report Date

6. Performing Organization Code505-33-22

8. Performing Organization Report No.

E-157810. Work Unit No.

11. Contract or Grant No.

13. Type of Report and Period Covered

Technical Memorandum14. Sponsoring Agency Code

15. Supplementary NottsPrepared for Let's Do It Ourselves: In Science and Technology sponsored by the NationalTechnical Association, Baltimore, Maryland, August 2-7, 1982.

16. Abstract

This Investigation was conducted to determine the effects of welding, selected surface coatingsand stress relieving on the thermal fatigue resistance of H- 13 Die Steel for aluminum diecasting dies. Eleven thermal fatigue specimens were utilized In this Investigation and arecatagorlzed as follows: control, stress relieved, welded, and surface treated (5 differentsurface treatments were evaluated). Stress relieving was conducted after each 5, 000 cycleInterval at 1050°F for three hours. Four thermal fatigue specimens were welded with H-13 ormaraging steel welding rods at ambient and elevated temperatures and subsequently, subjectedto different post- weld heat treatments. The surface treatments used are proprietary but aredescribed In generic terms. Crack patterns were examined at 5,000, 10,000, and 15,000 cycles.The thermal fatigue resistance Is expressed by two crack parameters which are the averagemaximum crack and the average cracked area. The results Indicated that a significantimprovement in thermal fatigue resistance over the control was obtained from the stress-relieving treatment. Small Improvements in thermal fatigue resistance were obtained from theH-13 welded specimens and from a salt bath nitrogen and carbon- surf ace treatment which isattributed to the alleviation of the residual stresses and the formation of a diffusion layer at thesurface. The other surface treatments and welded specimens either did not affect or had adetrimental influence on the thermal fatigue properties of the H-13 die steel.

17. Key Words (Suggested by Author ( i l lThermal fatigueH-13 die steelDie casting diesHeat checks

18. Distribution Statement

Unclassified - unlimitedSTAR Category 26

19. Security Oassif. (of this report) 20. Security Qassif. (of this page)

Unclassified Unclassified21. No. of Pages 22. Price*

1 For sale by the National Technical Information Service. Springfield. Virginia 22161